A method and computing device for predicting power load in remote areas

Abstract

The present invention discloses a method, computing device and storage medium for power load forecasting in remote areas. The method comprises: preprocessing historical power load data in remote areas to obtain load data to be input; inputting the load data to be input into a trained neural network model to output first predicted load data; obtaining second predicted load data through a comprehensive forecasting method; and combining the first predicted load data with the second predicted load data for forecasting to determine final predicted load data.

Description

Technical Field

[0001] The present invention relates to the field of energy and electricity, and in particular to a power load forecasting method, computing equipment and storage medium suitable for remote areas. Background Art

[0002] The development of renewable energy, particularly wind turbines and photovoltaic technologies, offers enormous potential for providing reliable electricity to remote areas. Optimizing grid design is essential for achieving lower costs and higher reliability. Because electricity supply is significantly affected by demand, load is one of the most influential factors in power system design.

[0003] Electricity load, as a crucial data input for grid planning and operational research, significantly impacts the performance of power grids in remote areas. Accurate load forecasting can improve the operational efficiency of hybrid energy systems. However, current solutions for electricity load forecasting primarily target industrial parks or cities, with less attention paid to remote areas, and their accuracy is limited. In reality, remote areas, with their vast expanses and complex and diverse terrain, are crucial for developing renewable energy and achieving carbon peak and carbon neutrality.

[0004] Therefore, a new power load forecasting method suitable for remote areas is needed for optimization processing. Summary of the Invention

[0005] To this end, the present invention provides a power load forecasting solution applicable to remote areas, in an effort to solve or at least alleviate the above problems.

[0006] According to one aspect of the present invention, a method for power load forecasting suitable for remote areas is provided, the method comprising the following steps: first, preprocessing historical power load data of the remote areas to obtain load data to be input; inputting the load data to be input into a trained neural network model to output first predicted load data; obtaining second predicted load data through a comprehensive forecasting method; and combining the first predicted load data with the second predicted load data for forecasting to determine final predicted load data.

[0007] Optionally, in the power load forecasting method applicable to remote areas according to the present invention, the step of preprocessing the historical power load data of the remote areas to obtain the load data to be input includes: matrixing the historical power load data of the remote areas to obtain matrix load data; performing feature selection on the matrix load data to generate characteristic load data; and standardizing the characteristic load data to obtain the load data to be input.

[0008] Optionally, in the power load forecasting method applicable to remote areas according to the present invention, the neural network model includes a first prediction layer and a second prediction layer, the input of the first prediction layer includes the load data to be input and the output of the second prediction layer, the output of the first prediction layer is the first predicted load data, and the input of the second prediction layer includes the load data to be input and the output of the first prediction layer.

[0009] Optionally, in the power load forecasting method applicable to remote areas according to the present invention, the first forecasting layer and the second forecasting layer are both multi-layer perceptron artificial neural networks.

[0010] Optionally, in the power load forecasting method applicable to remote areas according to the present invention, the step of obtaining the second forecast load data through a comprehensive forecasting method includes: conducting a comprehensive analysis of users in remote areas to predict the user's reported load; predicting the natural growth load through the natural growth rate method; and taking the sum of the user's reported load and the natural growth load as the second forecast load data.

[0011] Optionally, in the power load forecasting method applicable to remote areas according to the present invention, a comprehensive analysis is conducted on users in remote areas to predict the user's reported load, including: conducting a comprehensive analysis on users in remote areas to obtain the expected reported capacity, the load end-period practical coefficient and the load stage coefficient; calculating the product of the expected reported capacity, the load end-period practical coefficient and the load stage coefficient as the user's reported load.

[0012] Optionally, in the power load forecasting method applicable to remote areas according to the present invention, the step of combining the first forecast load data and the second forecast load data for forecasting to determine the final forecast load data includes: calculating the absolute difference between the first forecast load data and the second forecast load data; and determining the final forecast load data based on the relationship between the absolute difference and a preset threshold.

[0013] Optionally, in the power load forecasting method applicable to remote areas according to the present invention, the step of determining the final forecast load data based on the size relationship between the absolute difference and the preset threshold includes: if the absolute difference is less than or equal to the preset threshold, the first forecast load data is used as the final forecast load data.

[0014] Optionally, in the power load forecasting method applicable to remote areas according to the present invention, the step of determining the final forecast load data based on the size relationship between the absolute difference and a preset threshold value includes: if the absolute difference is greater than the preset threshold value, performing weighted summation on the first forecast load data and the second forecast load data to determine the final forecast load data.

[0015] Optionally, in the power load forecasting method applicable to remote areas according to the present invention, the step of performing weighted summation on the first predicted load data and the second predicted load data to determine the final predicted load data includes: generating a variance expression of the final predicted load data based on a first prediction error of the first predicted load data and a second prediction error of the second predicted load data; introducing a Lagrange multiplier to the variance expression to find a minimum value to solve the first weight coefficient of the first predicted load data and the second weight coefficient of the second predicted load data; and calculating the weighted sum of the first predicted load data and the second predicted load data as the final predicted load data through the first weight coefficient and the second weight coefficient.

[0016] According to another aspect of the present invention, a computing device is provided, comprising: at least one processor; and a memory storing program instructions, wherein the program instructions are configured to be suitable for execution by the at least one processor, and the program instructions include instructions for executing the power load forecasting method applicable to remote areas as described above.

[0017] According to another aspect of the present invention, a readable storage medium storing program instructions is provided. When the program instructions are read and executed by a computing device, the computing device executes the above-mentioned power load forecasting method applicable to remote areas.

[0018] According to the power load forecasting scheme for remote areas of the present invention, historical power load data for remote areas is preprocessed to obtain load data to be input, the load data to be input is input into a trained neural network model to output first predicted load data, second predicted load data is obtained through a comprehensive forecasting method, and the first predicted load data and the second predicted load data are combined for forecasting to determine the final predicted load data. In the above technical scheme, neural network forecasting and comprehensive forecasting methods are combined. The comprehensive forecasting method can adopt a combination of the natural growth rate method and the user comprehensive analysis method, which can more accurately forecast the power load in remote areas. Accurate load forecasting can improve the operating efficiency of the hybrid energy system, thereby better supporting the power grid planning in remote areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] To achieve the above and related purposes, certain illustrative aspects are described herein in conjunction with the following description and accompanying drawings, which indicate various ways in which the principles disclosed herein may be practiced, and all aspects and their equivalents are intended to fall within the scope of the claimed subject matter. The above and other objects, features, and advantages of the present disclosure will become more apparent by reading the following detailed description in conjunction with the accompanying drawings. Throughout this disclosure, the same reference numerals generally refer to the same parts or elements.

[0020] Figure 1shows a structural block diagram of a computing device 100 according to one embodiment of the present invention;

[0021] Figure 2 A flowchart of a method 200 for predicting power load in remote areas according to an embodiment of the present invention is shown; and

[0022] Figure 3 A schematic structural diagram of a neural network model according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0023] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0024] Figure 1 FIG. 1 shows a structural block diagram of a computing device 100 according to an embodiment of the present invention.

[0025] like Figure 1 As shown, in a basic configuration 102, computing device 100 typically includes system memory 106 and one or more processors 104. A memory bus 108 may be used for communication between processor 104 and system memory 106.

[0026] Depending on the desired configuration, the processor 104 can be any type of processor, including but not limited to: a microprocessor (UP), a microcontroller (UC), a digital signal processing unit (DSP), or any combination thereof. The processor 104 can include one or more levels of cache, such as a level 1 cache 110 and a level 2 cache 112, a processor core 114, and registers 116. An example processor core 114 can include an arithmetic logic unit (ALU), a floating point unit (FPU), a digital signal processing core (DSP core), or any combination thereof. An example memory controller 118 can be used with the processor 104, or in some implementations, the memory controller 118 can be an internal part of the processor 104.

[0027] Depending on the desired configuration, system memory 106 can be any type of memory, including but not limited to volatile memory (such as RAM), non-volatile memory (such as ROM, flash memory, etc.), or any combination thereof. System memory 106 can include an operating system 120, one or more applications 122, and program data 124. In some embodiments, application 122 can be arranged to execute instructions on the operating system by one or more processors 104 using program data 124.

[0028] Computing device 100 also includes storage 132 , which includes removable storage 136 and non-removable storage 138 .

[0029] The computing device 100 may also include a storage interface bus 134. The storage interface bus 134 enables communication from storage devices 132 (e.g., removable storage 136 and non-removable storage 138) to the basic configuration 102 via the bus / interface controller 130. At least a portion of the operating system 120, applications 122, and program data 124 may be stored on the removable storage 136 and / or the non-removable storage 138 and loaded into the system memory 106 via the storage interface bus 134 when the computing device 100 is powered on or when the application 122 is to be executed, and executed by the one or more processors 104.

[0030] The computing device 100 may also include an interface bus 140 that facilitates communication from various interface devices (e.g., output devices 142, peripheral interfaces 144, and communication devices 146) to the basic configuration 102 via the bus / interface controller 130. Example output devices 142 include a graphics processing unit 148 and an audio processing unit 150. These can be configured to facilitate communication with various external devices such as a display or speakers via one or more A / V ports 152. Example peripheral interfaces 144 may include a serial interface controller 154 and a parallel interface controller 156, which can be configured to facilitate communication with external devices such as input devices (e.g., a keyboard, mouse, pen, voice input device, touch input device) or other peripherals (e.g., a printer, scanner, etc.) via one or more I / O ports 158. Example communication devices 146 may include a network controller 160, which can be arranged to facilitate communication with one or more other computing devices 162 via a network communication link via one or more communication ports 164.

[0031] A network communication link can be an example of a communication medium. Communication media can generally be embodied as computer-readable instructions, data structures, program modules in a modulated data signal such as a carrier wave or other transmission mechanism, and can include any information delivery medium. A "modulated data signal" can be a signal in which one or more of a data set or a change therein can be carried out in a manner that encodes information in the signal. As non-limiting examples, communication media can include wired media such as a wired network or a dedicated line network, and various wireless media such as sound, radio frequency (RF), microwave, infrared (IR) or other wireless media. The term computer-readable medium as used herein can include both storage media and communication media.

[0032] The computing device 100 can be implemented as a personal computer including desktop and notebook computer configurations. Of course, the computing device 100 can also be implemented as part of a small-sized portable (or mobile) electronic device, such as a cellular phone, a digital camera, a personal digital assistant (PDA), a personal media player device, a wireless network browsing device, a personal head-mounted device, an application-specific device, or a hybrid device that can include any of the above functions. It can even be implemented as a server, such as a file server, a database server, an application server, and a web server. The embodiments of the present invention are not limited to this.

[0033] In an embodiment according to the present invention, the computing device 100 is configured to execute the method 200 for power load forecasting in remote areas according to the present invention. The application 122 deployed on the operating system includes multiple program instructions for executing the method 200. These program instructions can instruct the processor 104 to execute the method 200 of the present invention, so that the computing device 200 can perform power load forecasting for remote areas by executing the method 200 of the present invention.

[0034] Figure 2 A flow chart of a method 200 for predicting power loads in remote areas according to an embodiment of the present invention is shown. The method 200 for predicting power loads in remote areas may be executed in a computing device (eg, the aforementioned computing device 100).

[0035] like Figure 2 As shown, the method 200 begins at step S210. In step S210, historical power load data of remote areas is pre-processed to obtain load data to be input.

[0036] Remote areas are located at the end of urban and rural areas, with relatively underdeveloped economies, populations, and production and construction. Due to geographical constraints, remote areas generally have small populations, maintaining generally stable growth. They also have dispersed populations, low population density, and relatively underdeveloped economies, often relying on traditional agricultural production.

[0037] According to one embodiment of the present invention, historical power load data in remote areas can be preprocessed in the following manner to obtain load data to be input: First, the historical power load data in remote areas is matrixed to obtain matrix load data, and then feature selection is performed on the matrix load data to generate characteristic load data, and the characteristic load data is standardized to obtain the load data to be input.

[0038] In this embodiment, the historical power load data is historical daily load data, and the matrix load data obtained after matrix processing can be expressed as follows:

[0039]

[0040] Among them, L represents the matrix load data, L d-D (D = 1, 2, ..., 1699) represents the load demand D days ago. Each row in the matrix can be considered a variable. When performing feature selection on matrix load data, the important variables are selected from these variables as feature data to generate characteristic load data.

[0041] According to one embodiment of the present invention, feature selection can be performed using statistical correlation and mutual information methods. These methods are not discussed here and will be discussed later in the section on neural network model training. Min-Max Normalization is used to perform a linear transformation on the characteristic load data, placing the result in the range [0, 1] to obtain the input load data.

[0042] Then, the process proceeds to step S220 , where the load data to be input is input into the trained neural network model to output first predicted load data. Figure 3 FIG. 1 shows a schematic diagram of the structure of a neural network model according to an embodiment of the present invention. Figure 3 As shown, the neural network model includes a first prediction layer and a second prediction layer. The input of the first prediction layer includes the load data to be input and the output of the second prediction layer. The output of the first prediction layer is the first predicted load data. The input of the second prediction layer includes the load data to be input and the output of the first prediction layer. Both the first prediction layer and the second prediction layer are Multi-Layer Perceptron-Artificial Neural Networks (MLP-ANNs).

[0043] For MLP-ANN, it generally includes three layers: input layer, hidden layer and output layer. The hidden layer is a layer with multiple neurons (i.e., nodes), and the number of neurons in the input layer and the output layer is usually determined according to the required input and output variables, respectively. In this embodiment, the number of neurons in the hidden layer is 12. All neurons in the input layer are summed and provide input for the hidden layer neurons. The output of the output layer gives the predicted load, that is, the load parameters for the next few days. Each neuron in MLP-ANN has the characteristics of receiving input, processing and providing output, and is assigned a weight vector to connect it to all neurons in the subsequent layers.

[0044] Before applying the neural network model to predictions, it needs to be trained on a training dataset. The iterative training process stops when the prediction error between the first and second prediction layers meets a certain standard. Otherwise, the training process continues. The training dataset includes 86% of the total data, and the remaining data serves as the test and validation datasets.

[0045] Before using the training dataset to train the neural network model, the training dataset needs to be preprocessed to obtain a standard dataset suitable for training. The method for obtaining this standard dataset can refer to the preprocessing process in step S210, that is, the training dataset is first matrixed to obtain a matrix dataset, and then feature selection is performed on the matrix dataset to generate a feature dataset. Finally, the feature dataset is normalized to obtain the standard dataset.

[0046] Statistical correlation and mutual information methods use feature selection to identify which variables influence the predicted value, thereby extracting these variables to reduce the dimensionality of the data. This is equivalent to first filtering the matrixed data using the feature selection process and then standardizing the filtered data to facilitate subsequent model training.

[0047] Statistical correlation is a concept in statistics that measures how closely the trend of a feature's value matches the trend of past actual values. The correlation coefficient is a number between -1 and 1. If there is no relationship between the feature's value and past values, the correlation coefficient is 0 or very low. As the strength of the relationship between the feature's value and past values ​​increases, the correlation coefficient also increases. A perfect fit results in a correlation coefficient of 1. Therefore, the higher the correlation coefficient, the higher the quality of the input used as an MLP-ANN. One advantage of this work is that it eliminates invalid input parameters, resulting in a smaller and simpler artificial neural network architecture.

[0048] For variables A and B, they have their own expected values ​​μ A and μ B , and the standard deviation σ A and σB , then the overall correlation coefficient corr(A,B) can be expressed as follows:

[0049]

[0050] Among them, cov represents covariance, E represents expected value, and the overall correlation coefficient corr(A,B) is mainly used to measure the degree of linear correlation.

[0051] Mutual information works similarly to statistical correlation, but uses a different approach. Mutual information is a method for selecting a small set of correlated features from a high-dimensional dataset. This process involves a series of estimates of the relationship between each pair of variables and between each variable and the class label. Mutual information-based predictions are widely used in the development of feature selection methods. The advantage of using mutual information to estimate the dependency between two variables is that it can measure not only linear or monotonic relationships, but also more complex nonlinear relationships.

[0052] The mutual information I(X; Y) of two random discrete variables X and Y can be expressed as follows:

[0053]

[0054] Where p(x,y) represents the joint probability distribution function of X and Y, and p(x) and p(y) represent the marginal probability distribution functions of X and Y, respectively. When X and Y are continuous random variables, the mutual information I(X;Y) is transformed from the summation in Equation (3) into the following double integral:

[0055]

[0056] Where p(x,y) represents the joint probability density function of X and Y, while p(x) and p(y) represent the marginal probability density functions of X and Y, respectively.

[0057] In step S230, second predicted load data is obtained through a comprehensive forecasting method. According to one embodiment of the present invention, the second predicted load data can be obtained by performing a comprehensive analysis of users in remote areas to predict their reported loads. The natural growth rate method is then used to predict the naturally increasing load. The sum of the user reported load and the naturally increasing load is used as the second predicted load data.

[0058] Because user development trends reflect macroeconomic trends, national and regional economic policies, the characteristics of the regional economy's industrial structure, and the region's periodic resource advantages (energy, minerals, land, transportation, water resources, etc.), the concentration of new users in a given region represents the hotspots and characteristics of that region's economic development and reflects the particular contradictions in the macroeconomic development process. Therefore, when predicting user load applications, it is necessary to first conduct a comprehensive analysis of users in remote areas based on user development information to obtain the estimated application capacity, the load end-period utility coefficient, and the load stage coefficient. The product of the estimated application capacity, the load end-period utility coefficient, and the load stage coefficient is then calculated as the user's reported load.

[0059] In this embodiment, the user reports the load P 报装负荷 The specific calculation formula is as follows:

[0060] P 报装负荷 =P 报装容量 ×λ 实用系数 ×β 阶段系数 (5)

[0061] Among them, P 报装容量 is the estimated capacity, λ 实用系数 is the load terminal utility coefficient, β 阶段系数 is the load stage factor.

[0062] Assuming that the load growth pattern is the same as that of historical data, the average proportional coefficient can be calculated to predict the development trend. Assuming that the current year is τ b , then the average peak power in the first quarter is The initial year is τ a , then the average peak power in the first quarter is Then, the average growth rate K of the average peak power from the initial year to the first quarter of the current year is:

[0063]

[0064] The average peak power in the first quarter of year m is thus predicted for:

[0065]

[0066] The above is the prediction principle of the natural growth rate method, with P 自然增长负荷 It represents the natural growth load predicted by the natural growth rate method. f1 and f2 represent the first predicted load data and the second predicted load data respectively. For the second predicted load data f2, its value is:

[0067] f2=P 自然增长负荷 +P 报装负荷 (8)

[0068] Finally, step S240 is executed to combine the first predicted load data and the second predicted load data to determine the final predicted load data. According to one embodiment of the present invention, the final predicted load data can be determined as follows: the absolute difference between the first predicted load data and the second predicted load data is first calculated, and then the final predicted load data is determined based on the relationship between the absolute difference and a preset threshold.

[0069] In this embodiment, if the absolute difference is less than or equal to the preset threshold, the first predicted load data is used as the final predicted load data; if the absolute difference is greater than the preset threshold, the first predicted load data and the second predicted load data are weightedly summed to determine the final predicted load data.

[0070] Let σ be the preset threshold (i.e. 2% of the arithmetic mean of the two prediction results). When the absolute difference between f1 and f2 is less than or equal to σ, the final predicted load data f = f1. When the absolute difference between f1 and f2 is greater than σ, it is necessary to use the weighted summation of f1 and f2 (i.e. the variance-covariance optimal combined prediction method) to determine the final predicted load data f.

[0071] According to one embodiment of the present invention, a weighted sum of the first predicted load data and the second predicted load data may be performed as follows to determine the final predicted load data. A variance expression for the final predicted load data is generated based on a first prediction error of the first predicted load data and a second prediction error of the second predicted load data. A Lagrange multiplier is introduced into the variance expression to find a minimum value, thereby solving for a first weight coefficient for the first predicted load data and a second weight coefficient for the second predicted load data. Using the first weight coefficient and the second weight coefficient, a weighted sum of the first predicted load data and the second predicted load data is calculated as the final predicted load data.

[0072] Let f=ω1f1+ω2f2, where ω1 is the first weight coefficient of the first predicted load data f1, and ω2 is the second weight coefficient of the second predicted load data f2.

[0073] Assuming that the first prediction error of the first predicted load data f1 is e1, and the second prediction error of the second predicted load data f2 is e2, and f is required to be unbiased, the prediction error e of f is:

[0074] e=ω1e1+ω2e2 (9)

[0075] Because e1 and e2 are independent of each other, the variance expression Var(e) of the final predicted load data f is:

[0076]

[0077] Among them, Var(e1) and Var(e2) represent the variance of f1 and f2 respectively.

[0078] Note Var(e i )=σ ii , where e i is the prediction result f i The prediction error, Var(e i ) represents f i The variance of i =1, introducing the Lagrange multiplier to find the minimum value of Var(e) yields:

[0079]

[0080]

[0081] Among them, ω i is the prediction result f i The weight coefficient, σ ii , σ 11 , σ 22 Respectively represent f i , the variance of f1, and f2.

[0082] e1 and e2 obey zero-mean normal distribution, σ ii Estimated value of It can be determined by the following formula:

[0083]

[0084] Among them, e it is the prediction result f i The prediction result error of the tth data, n is the number of historical data.

[0085] From this, the estimated formulas for the first weight coefficient and the second weight coefficient can be obtained as follows:

[0086]

[0087] in, ω i The estimated value of jt is the prediction result f j The prediction result error of the tth data.

[0088] Right now, is the estimated value of the first weight coefficient ω1, is the estimated value of the second weight coefficient ω2, and then we can Instead of ω1, The final predicted load data f is calculated instead of ω2.

[0089] According to an embodiment of the present invention, a power load forecasting scheme for remote areas preprocesses historical power load data in remote areas to obtain load data to be input, inputs the load data to be input into a trained neural network model to output first predicted load data, obtains second predicted load data through a comprehensive forecasting method, and combines the first predicted load data with the second predicted load data for forecasting to determine final predicted load data. In the above technical solution, neural network forecasting and comprehensive forecasting methods are combined. The comprehensive forecasting method can adopt a combination of the natural growth rate method and the user comprehensive analysis method, which can more accurately forecast the power load in remote areas. Accurate load forecasting can improve the operating efficiency of the hybrid energy system, thereby better supporting power grid planning in remote areas.

[0090] A9. A method as described in A7 or A8, wherein the step of determining the final predicted load data based on the relationship between the absolute difference and a preset threshold value includes: if the absolute difference is greater than the preset threshold value, performing a weighted summation on the first predicted load data and the second predicted load data to determine the final predicted load data.

[0091] A10. A method as described in A9, wherein the step of performing weighted summation on the first predicted load data and the second predicted load data to determine the final predicted load data includes: generating a variance expression for the final predicted load data based on a first prediction error of the first predicted load data and a second prediction error of the second predicted load data; introducing a Lagrange multiplier to the variance expression to find a minimum value to solve a first weight coefficient of the first predicted load data and a second weight coefficient of the second predicted load data; and calculating the weighted sum of the first predicted load data and the second predicted load data as the final predicted load data using the first weight coefficient and the second weight coefficient.

[0092] The various techniques described herein may be implemented in conjunction with hardware or software, or a combination thereof. Thus, the methods and apparatus of the present invention, or certain aspects or portions of the methods and apparatus of the present invention, may be implemented in the form of program codes (i.e., instructions) embedded in a tangible medium, such as a removable hard disk, a USB flash drive, a floppy disk, a CD-ROM, or any other machine-readable storage medium, wherein when the program is loaded into a machine such as a computer and executed by the machine, the machine becomes an apparatus for practicing the present invention.

[0093] When the program code is executed on a programmable computer, the computing device generally includes a processor, a storage medium readable by the processor (including volatile and non-volatile memory and / or storage elements), at least one input device, and at least one output device. The memory is configured to store the program code, and the processor is configured to execute the power load forecasting method for remote areas of the present invention according to the instructions in the program code stored in the memory.

[0094] By way of example and not limitation, readable media include readable storage media and communication media. Readable storage media store information such as computer-readable instructions, data structures, program modules, or other data. Communication media typically embody computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and include any information delivery medium. Combinations of any of the above are also included within the scope of readable media.

[0095] In the description provided herein, the algorithms and displays are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems may also be used in conjunction with the examples of the present invention. Based on the above description, it is apparent that the structure required for constructing such systems is well understood. In addition, the present invention is not directed to any specific programming language. It should be understood that various programming languages ​​may be utilized to implement the present invention described herein, and the description of specific languages ​​above is provided for the purpose of disclosing the preferred embodiment of the present invention.

[0096] In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures, and techniques are not shown in detail so as not to obscure the understanding of this description.

[0097] Similarly, it should be understood that in order to streamline the present disclosure and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, this disclosed method should not be interpreted as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim. Rather, as reflected in the claims below, inventive aspects lie in less than all the features of the individual embodiments disclosed above. Accordingly, the claims that follow the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of the invention.

[0098] Those skilled in the art will appreciate that the modules, units, or components of the devices in the examples disclosed herein may be arranged in the device described in the embodiment, or alternatively may be located in one or more devices different from the devices in the examples. The modules in the foregoing examples may be combined into one module or further divided into multiple submodules.

[0099] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and in addition may be divided into multiple submodules or subunits or subcomponents. All features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed herein may be combined in any combination, except that at least some of such features and / or processes or units are mutually exclusive. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature providing the same, equivalent or similar purpose.

[0100] Furthermore, those skilled in the art will appreciate that although some embodiments described herein include certain features included in other embodiments but not other features, combinations of features from different embodiments are intended to be within the scope of the present invention and to form different embodiments. For example, in the claims below, any of the claimed embodiments may be used in any combination.

[0101] In addition, some of the embodiments are described herein as methods or combinations of method elements that can be implemented by a processor of a computer system or by other devices that perform the functions described. Thus, a processor having the necessary instructions for implementing the method or method element forms a device for implementing the method or method element. Furthermore, the elements described herein of the device embodiments are examples of devices for implementing the functions performed by the elements for the purpose of implementing the invention.

[0102] As used herein, unless otherwise specified, the use of ordinal numbers "first," "second," "third," etc. to describe common objects merely indicates that different instances of similar objects are involved and are not intended to imply that the objects so described must have a given order in time, space, ranking, or in any other manner.

[0103] Although the present invention has been described with respect to a limited number of embodiments, it will be apparent to those skilled in the art, having benefit of the foregoing description, that other embodiments are contemplated within the scope of the invention thus described. Furthermore, it should be noted that the language used in this specification has been selected primarily for readability and didactic purposes, rather than for the purpose of explaining or limiting the subject matter of the present invention. Consequently, many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the appended claims. The disclosure of the present invention is intended to be illustrative rather than restrictive of the scope of the invention, which is defined by the appended claims.

Claims

1. A method for predicting power load in remote areas, comprising: Preprocessing historical power load data of remote areas to obtain load data to be input, including: matrixing the historical power load data of remote areas to obtain matrix load data, performing feature selection on the matrix load data through statistical correlation and mutual information to generate characteristic load data, and standardizing the characteristic load data to obtain the load data to be input; Inputting the load data to be input into a trained neural network model to output first predicted load data, wherein the neural network model includes a first prediction layer and a second prediction layer, the input of the first prediction layer includes the load data to be input and the output of the second prediction layer, the output of the first prediction layer is the first predicted load data, and the input of the second prediction layer includes the load data to be input and the output of the first prediction layer; Obtaining second predicted load data by a comprehensive prediction method, including: performing a comprehensive analysis on users in the remote area to predict the user's reported load, predicting the natural growth load by a natural growth rate method, and taking the sum of the user's reported load and the natural growth load as the second predicted load data; The first predicted load data and the second predicted load data are combined and predicted to determine the final predicted load data, including: calculating the absolute difference between the first predicted load data and the second predicted load data, and determining the final predicted load data based on the relationship between the absolute difference and a preset threshold, wherein if the absolute difference is less than or equal to the preset threshold, the first predicted load data is used as the final predicted load data; if the absolute difference is greater than the preset threshold, the first predicted load data and the second predicted load data are weightedly summed to determine the final predicted load data.

2. The method according to claim 1, wherein the first prediction layer and the second prediction layer are both multi-layer perceptron artificial neural networks.

3. The method according to claim 1 or 2, wherein The step of comprehensively analyzing the users in the remote areas to predict the user's reported load includes: Conduct a comprehensive analysis of users in the remote areas to obtain the estimated installed capacity, load terminal utility coefficient, and load stage coefficient; The product of the estimated reported capacity, the load terminal utility coefficient and the load stage coefficient is calculated as the user reported load.

4. The method according to claim 1, wherein The step of performing weighted summation on the first predicted load data and the second predicted load data to determine final predicted load data comprises: generating a variance expression for final predicted load data according to a first prediction error of the first predicted load data and a second prediction error of the second predicted load data; Introducing a Lagrange multiplier into the variance expression to find a minimum value, so as to solve a first weight coefficient of the first predicted load data and a second weight coefficient of the second predicted load data; The weighted sum of the first predicted load data and the second predicted load data is calculated as final predicted load data using the first weight coefficient and the second weight coefficient.

5. A computing device comprising: at least one processor; as well as A memory storing program instructions, wherein the program instructions are configured to be executed by the at least one processor, and the program instructions include instructions for executing the method according to any one of claims 1 to 4.

6. A readable storage medium storing program instructions, wherein when the program instructions are read and executed by a computing device, the computing device is caused to execute the method according to any one of claims 1 to 4.

Images